□ ORIGINAL ARTICLE □

Preventive Effect of Sulfamethoxasole-trimethoprim on Pneumocystis jiroveci Pneumonia in Patients with Interstitial Pneumonia

Tatsuji Enomoto 1,2,ArataAzuma2, Aki Matsumoto 2, Takahito Nei 2, Kazue Fujita 2, Kumiko Hattori 1,2, Yoshinobu Saito 2, Shinji Abe 2, Jiro Usuki 2 and Shoji Kudoh 2

Abstract

Background Pneumocystis jiroveci pneumonia (PCP) is a potentially fatal complication in interstitial pneu- monia patients receiving glucocorticoid therapy. Prophylaxis of PCP during glucocorticoid therapy is an im- portant issue in the treatment of interstitial pneumonia. Objective We evaluated the prophylactic effect of sulfamethoxasole-trimethoprim (TMP-SMX) in interstitial pneumonia patients receiving glucocorticoids. Methods We retrospectively analyzed 74 interstitial pneumonia patients who received glucocorticoid ther- apy. Results Seven of the 74 patients developed PCP. At the time of diagnosis of PCP, the mean duration of glucocorticoid therapy was 71 days and the mean daily dose of prednisolone was 37 mg. Among the 7 pa- tients, the circulating CD4+ lymphocyte count was 370/μl on average and it was over 200/μl in 3 cases. The PCP patients showed a significant reduction of the lymphocyte count at 4 weeks after initiation of steroid therapy. None of the patients who received prophylactic TMP-SMX therapy developed PCP even if the CD4+ lymphocyte count was less than 200/μl. Conclusion Interstitial pneumonia patients receiving glucocorticoid therapy can benefit from TMP-SMX prophylaxis against PCP. Development of PCP cannot be ruled out in patients with a CD4+ lymphocyte count of greater than 200/μl.

Key words: interstitial pneumonia, pneumocystis pneumonia, circulating CD4+ lymphocyte count, prophy- laxis with trimethoprim-sulfamethoxazole, beta-glucan

(DOI: 10.2169/internalmedicine.47.0402)

stitial pneumonia, prophylactic intervention against infection Introduction may lead to a better outcome of glucocorticoid/immunosup- pressant therapy. Although glucocorticoids and immunosuppressants are In this study, we retrospectively evaluated the efficacy of widely used for the treatment of idiopathic interstitial pneu- trimethoprim-sulfamethoxazole (TMP-SMX) for prophylaxis monia (1), as well as that associated with connective tissue of PCP during glucocorticoid and/or immunosuppressant diseases and vascular disorders (2), glucocorticoid/immuno- therapy for interstitial pneumonia. suppressant therapy has adverse effects such as the oc- curence of opportunistic infections in the lungs, including Materials and Methods Pneumocystis jiroveci pneumonia (PCP), other fungal infec- tions, and cytomegalovirus infection (3, 4). Because of the This study was a retrospective analysis of 74 patients with risk of opportunistic infection during the treatment of inter- interstitial pneumonia (IP) who were admitted to the Depart-

１Department of Respiratory Medicine, Tokyo Metropolitan Hiroo General Hospital, Tokyo and ２Division of Pulmonary Diseases, Infection, and Oncology, Nippon Medical School, Tokyo Received for publication June 23, 2007; Accepted for publication September 8, 2007 Correspondence to Dr. Arata Azuma, [email protected]

15 DOI: 10.2169/internalmedicine.47.0402

Table 1. Clinical Features of Patients with Pneumocystis jiroveci Pneumonia

ment of Respiratory Medicine at Nippon Medical School matomyositis, 1 with systemic sclerosis, 3 with Sjogren syn- Hospital, and who received more than 0.5 mg/kg of prednis- drome, and 3 with microscopic polyangitis), pneumoconiosis olone with or without immunosuppressants for more than in 1 patient, and chronic hypersensitivity pneumonia in 1 three weeks between July 1997 and September 2003. We patient. evaluated the following clinical parameters: age, sex, under- Fourteen (18.9%) of the 74 patients developed opportunis- lying disease, body mass index, smoking status, diabetes tic pulmonary infection: 7 developed PCP, 3 had cytomega- mellitus, the dose and duration of glucocorticoid therapy, lovirus pneumonia, 1 developed aspergillus pneumonia, 3 use of immunosuppressants, the white blood cell count be- had expanded MRSA pneumonia, and 3 developed sepsis. fore and 4 weeks after the initiation of glucocorticoid ther- The ten patients without PCP infection were excluded from apy, the results of pulmonary function tests before therapy, this study. We only evaluated the first infection in each pa- and the response to prophylaxis with trimethoprim- tient. All 7 patients showed high-resolution computed to- sulfamethoxazole (TMP-SMX). In the IP patients who de- mography findings consistent with PCP. Two patients were veloped PCP, we also evaluated the white blood cell count, positive for pneumocystis on PCR of sputum samples, two circulating CD4+ lymphocyte count, and plasma level of had positive findings on microscopic examination of BALF beta-D-glucan at the time of diagnosis of PCP. samples, and five were positive for pneumocystis by PCR of The diagnostic criteria for PCP were: 1) appropriate find- BALF samples (including 2 patients who were positive on ings on high-resolution computed tomography, and 2) detec- both examinations). tion of Pneumocystis jiroveci by microscopic examination or We compared the 7 patients who developed PCP (PCP by a positive polymerase chain reaction (PCR) assay of spu- group) with the 57 patients who did not develop any oppor- tum and bronchoalveolar lavage fluid (BALF) samples (5). tunistic infection (control group) (Table 1). Age, sex, body Sputum and BALF samples were stained with Papanicolaou mass index, smoking status, results of pulmonary function stain, May-Giemsa stain, and Grocott’s stain for microscopic tests, presence of diabetes mellitus, use of steroid pulse ther- examination. apy, and use of immunosuppressants did not differ signifi- Differences between pairs of groups were examined for cantly between the two groups. However, the initial dose of statistical significance with the χ2 test or one-factor ANOVA. glucocorticoids was significantly different between the PCP Differences among three groups were examined for statisti- group and the control group patients with or without TMP- cal significance with repeated measures ANOVA. SMX prophylax (trimethoprim 80 mg-sulfamethoxazole 400 mg/day). In the control group, 15 of the 57 patients received Results prophylactic therapy with TMP-SMX, while none of the pa- tients in the PCP group received TMP-SMX prophylaxis. The seventy-four patients with interstitial pneumonia com- The white blood cell count and neutrophil count before prised 36 men and 38 women with a mean age of 64 years glucocorticoid therapy were significantly different between (range: 23-85 years). The underlying diseases were idi- the PCP and control groups. However, the lymphocyte count opathic pulmonary fibrosis in 20 patients, idiopathic intersti- four weeks after initiation of glucocorticoid therapy was not tial pneumonias other than idiopathic pulmonary fibrosis in significantly lower in the PCP group than in the control 24 patients, interstitial pneumonia related to collagen vascu- group (Table 2). Despite this, changes of the lymphocyte lar disease in 28 patients (5 with rheumatoid arthritis, 2 with count were significantly different among the three groups systemic lupus erythematosus, 14 with polymyositis/der- i.e., the PCP group, control patients with prophylaxis, and

16 DOI: 10.2169/internalmedicine.47.0402

Table 2. White Blood Cell Subsets in Patients with Pneumocystis jiroveci Pneumonia

Table 3. Cases of Pneumocystis jiroveci Pneumonia

control patients without prophylaxis (p<0.001 by repeated Table 4. Laboratory Data at the Time of Diagnosis of PCP measures ANOVA). The 7 patients who developed PCP had a mean age of 64 years (range: 55-76 years) and they consisted of three men and four women. Their underlying diagnoses were nonspe- cific interstitial pneumonia (n=2), interstitial pneumonia re- lated to polymyositis/dermatomyositis (n=3), interstitial pneumonia related to systemic sclerosis (n=1), and intersti- tial pneumonia related to microscopic polyangitis (n=1) (Ta- ble 3). The initial dose of glucocorticoid (prednisolone) was 30-100 mg/day, the average initial dose of prednisolone was 55 mg/day, and the cumulative dose was 1.530-16.859 g, and the mean cumulative dose was 5.551 g. The duration of glucocorticoid therapy was 38-175 days, with a mean dura- tion of 71 days. The dose of glucocorticoid at the time of The PCP group had a lower lymphocyte count (mean: diagnosis of PCP was 25-55 mg/day, with a mean dose of 1,133/μl) and a lower circulating CD4+ lymphocyte count 37 mg/day. In all 7 cases, PCP occurred after the acute (370/μl) in 5 of 7 patients. However, three patients who had phase of interstitial pneumonia. Three patients were receiv- a CD4+ lymphocyte count of greater than 200/μl developed ing immunosuppressant therapy at the time of diagnosis of PCP. Six of the seven PCP patients had a high beta-D- PCP (two patients were receiving cyclosporine at 150 mg/ glucan level (range: 66.7-1,320 pg/ml, mean: 360 pg/ml). day, and one patient was receiving cyclophosphamide at 50 The three patients who had a CD4+ lymphocyte count of mg/day). greater than 200/μl and developed PCP are presented below

17 DOI: 10.2169/internalmedicine.47.0402

Figure 1. In ‘c’ and ‘d’ patchy, bilateral ground-glass opacity is seen, whereas it is absent prior to treatment, ‘a’ and ‘b’, respectively.

(Tables 3, 4). prednisolone (40 mg/day). Four weeks after the beginning of A 69-year-old man (case No. 2) with microscopic poly- therapy, the dose was gradually tapered and the patient angitis and usual interstitial pneumonia was treated with showed a good response. Thirty-eight days after the begin- steroid pulse therapy (methylprednisolone at 1,000 mg for 3 ning of glucocorticoid therapy (on prednisolone at 30 mg/ days) and with prednisolone (50 mg daily). Four weeks after day; cumulative prednisolone dose: 1,700 mg), the patient the beginning of therapy, the dose of prednisolone was developed a high fever and hypoxia. Her white blood cell gradually tapered and the patient showed a good clinical re- count was 9,000/μl, lymphocyte count was 630/μl, CD4+ sponse. Forty-nine days after the start of glucocorticoid ther- lymphocyte count was 273/μl, and beta-D-glucan level was apy (cumulative dose: 4,165 mg), the patient developed a 68.2 pg/ml. The patient developed acute respiratory failure high fever and hypoxia. The white blood cell count was and required mechanical ventilation. At that time, high- 13,500/μl, lymphocyte count was 1,957/μl, CD4+ lympho- resolution computed tomography showed findings consistent cyte count was 836/μl, and beta-D-glucan level was 275 pg/ with PCP, and her BALF sample was positive for pneumo- ml. At that time, high-resolution computed tomography re- cystis by PCR. We diagnosed PCP. vealed patchy bilateral ground glass opacities in addition to honeycomb cysts (Fig. 1). Bronchoalveolar lavage fluid was Discussion positive for pneumocystis by PCR. We diagnosed PCP. A 55-year-old woman (case No. 4) with systemic sclero- Glucocorticoid therapy is associated with many adverse sis and usual interstitial pneumonia was treated with predni- effects. In particular, opportunistic infection is one of the solone (30 mg/day) and cyclophosphamide (50 mg/day). potentially fatal adverse effects. Stuck et al (4) studied the Four weeks after the beginning of therapy, the dose of pred- association between corticosteroid therapy and subsequent nisolone was gradually tapered and the patient showed a infections by pooling data from 71 controlled clinical trials, good clinical response. Fifty-eight days after the beginning and found that the rates of both lethal and nonlethal compli- of glucocorticoid therapy, she was on prednisolone at 20 cations were increased in patients who were treated with mg/day and cyclophosphamide at 50 mg/day (cumulative corticosteroids, with the relative risk being 2.6 and 1.6, re- prednisolone dose: 1,530 mg). At this time, the patient de- spectively. Takabayashi et al (6) reported that 28 (3.1%) of veloped a high fever and hypoxia. Her white blood cell 761 autoimmune disease patients who were treated with glu- count was 17,500/μl, lymphocyte count was 1,487/μl, CD4+ cocorticoid therapy developed pulmonary opportunistic in- lymphocyte count was 537/μl, and beta-D-glucan level was fection and 9 patients (1.2%) developed PCP. 66.7 pg/ml. High-resolution computed tomography showed Pneumocystis carinii pneumonia is a major cause of ill- patchy bilateral ground-glass opacities in addition to honey- ness and death among persons with an impaired immune comb cysts. Microscopic examination of BALF revealed system. Pneumocystis organisms from different host species Pneumocystis jiroveci, and the BALF sample was also posi- have very different DNA sequences. In recognition of its ge- tive for pneumocystis on PCR. We diagnosed PCP. netic and functional distinctness, the organism that causes A 76-year-old woman (case No. 5) with idiopathic non- human Pneumocystis pneumonia has been named Pneumo- specific interstitial pneumonia was started on therapy with cystis jiroveci (7).

18 DOI: 10.2169/internalmedicine.47.0402

In the present study, the rate of opportunistic pulmonary time of diagnosis of PCP. These results suggest that the de- infections among patients with interstitial pneumonia treated velopment of PCP after the start of glucocorticoid therapy is with glucocorticoid therapy was 23.0% (17/74), and that of based on changes that occur prior to therapy (Table 2). PCP was 9.5% (7/74). The PCP rate was higher than that The Japanese Respiratory Society Guidelines for the Man- shown in a previous study (6). One reason for the difference agement of Hospital-Acquired Pneumonia (11) state that in is thought to be improvement of diagnostic methods. An- immunodeficient patients, the risk of opportunistic pulmo- other reason may be that CD4+ lymphocytes are important nary infection (including PCP) increases when the CD4+ for host defenses against PCP (8). However, interstitial lymphocyte count falls below 200/μl. In the present study, at pneumonia patients who are treated with glucocorticoid ther- the time of diagnosis of PCP, the white blood cell count was apy not only have CD4+ lymphocyte depletion but are also not low (12,686/μl), but the lymphocyte count and CD4+ susceptible to pulmonary infection due to structural abnor- lymphocyte counts were low (1,133/μl and 370/μl on aver- malities of the lungs. age, respectively). However, three patients had a CD4+ lym- In the present patients who developed PCP, the duration phocyte count of greater than 200/μl at that time. These re- of glucocorticoid therapy was 71 days on average and the sults suggest that interstitial pneumonia patients who are re- mean dose of prednisolone at the time of diagnosis was 37 ceiving glucocorticoid therapy may be differently immuno- mg/day. These results were similar to those previously re- compromised than HIV-positive patients, and that their ported (6). The initial dose of prednisolone was significantly CD4+ lymphocytes may be dysfunctional because of posi- higher in the PCP group than in the control group, suggest- tive and negative modulation of the expression of several ing that a higher dose of glucocorticoids predisposes to this genes involved in inflammatory and immune responses by disease. The period of two to three months after the begin- glucocorticoid therapy (2). ning of glucocorticoid therapy is often the stable phase after Saito et al (12) evaluated 29 patients with various connec- the acute phase of interstitial pneumonia, and we must be tive tissue diseases, and reported that the prevalence of PCP aware of the possibility of PCP. was especially high among patients receiving >30 mg/day It was reported that in patients with human immunodefi- prednisolone with or without another immunosuppressant. ciency virus (HIV), the circulating CD4+ lymphocyte count Takabayashi et al (6) also reported that none of the patients was less than 200/μl before 46 of 49 episodes of PCP and receiving <30 mg/day prednisolone developed PCP. that patients with a CD4 count below 200/μl are likely to In the present study, none of the seven patients who de- benefit from pneumocystis prophylaxis (9). Gluck et al (10) veloped PCP were on prophylactic TMP-SMX therapy. The studied non-HIV patients receiving immunosuppressive ther- 15 patients who received prophylactic TMP-SMX therapy apy and reported that the CD4+ lymphocyte count was less did not develop PCP, even though the initial dose of predni- than 200/μl in all seven patients who developed PCP. They solone was 9 mg/day higher than in patients without TMP- concluded that measuring the CD4+ lymphocyte count was SMX prophylaxis from the control group (Table 1). Thus, helpful for determining the risk of PCP not only in HIV- interstitial pneumonia patients who are treated with gluco- positive patients, but also in patients receiving immunosup- corticoids may benefit from TMP-SMX prophylaxis against pressive therapy (10). In the present study, PCP occurred in PCP. A prophylactic effect of TMP-SMX against PCP was patients with structural abnormalities of the lungs, but Gluck originally reported in cancer patients (13). Sepkowitz et al et al reported on patients with collagen vascular diseases (14, 15) recommended that prophylaxis might be given to and hematological diseases that are generally not associated patients who receive glucocorticoid therapy for longer than with lung tissue distortion. This leads to a critical difference 4 weeks at a level equivalent to 20 mg of prednisone daily. in the risk of PCP. None of the present patients with idi- As a result of this investigation, we recommend that prophy- opathic pulmonary fibrosis developed PCP, whereas 5 out of laxis with TMP-SMX should be given if the lymphocyte 28 patients who had interstitial pneumonia associated with count prior to corticosteroid therapy is low (!1,000/μl) and collagen vascular disease developed PCP. Our collagen vas- if the lymphocyte count falls (even if it remains over 1,000/ cular disease patients all had nonspecific interstitial pneumo- μl) after starting corticosteroid therapy in patients with inter- nia. Differences of pathological findings were not signifi- stitial pneumonia. The present PCP patients received higher cantly associated with the risk of PCP in this study (data not corticosteroid doses than the controls, which would have in- shown). Pathological distortion may make the lungs more fluenced the lymphocyte count. susceptible to PCP, but the type of damage may be less im- β-D-glucan is one of the major components of the cyst portant. For this reason, PCP can develop even if the lym- wall in human Pneumocystis infection, and is reported to be phocyte count remains over 1,000/μl. We consider that the detectable in serum from patients with PCP. It serves as a risk factors for PCP not only include the CD4+ lymphocyte practical serological marker for monitoring the disease dur- count, but also structural damage to the lungs and changes ing treatment (16, 17). In the present study, all PCP patients of the lymphocyte count during treatment with glucocorti- who were evaluated showed a high serum β-D-glucan level, coids and/or immunosuppressants. The neutrophil count was so β-D-glucan is thought to be a good serological marker increased in the PCP group prior to glucocorticoid therapy, for PCP. although bacterial infection was not clinically evident at the In conclusion, we found that patients who had interstitial

19 DOI: 10.2169/internalmedicine.47.0402 pneumonia treated with glucocorticoids and/or immunosup- pneumonia patients receiving glucocorticoid/immunosupres- pressants could benefit from TMP-SMX prophylaxis against sant therapy. PCP. A CD4+ lymphocyte count of greater than 200/μl did not guarantee the prevention of PCP in patients with intersti- Acknowledgement tial pneumonia. The authors thank Dr. Seiichi Nakamura, Dr. Yasuhiro However, a prospective study is needed to confirm the Shibuya, and Dr. Takehumi Saito for suitable advice on the prophylactic effect of TMP-SMX against PCP in interstitial evaluation of statistical analysis.

References

1. American Thoracic Society and European Respiratory Society: apy. Infection 28: 227-230, 2000. Idiopathic pulmonary fibrosis: diagnosis and treatment: Interna- 11. The Committee for Japanese Respiratory Society guideline in the tional concensus statement. Am J Respir Crit Care Med 161: 646- management of respiratory infection, Pneumonia in immunosup- 664, 2000. pressed patients. The Japanese Respiratory Society Guidelines for 2. Boumpas DT, Chrousos GP, Wilder RL, Cupps TR, Balow JE. Management of Hospital-acquired Pneumonia. The Japanese Glucocorticoid therapy for immune-mediated diseases: basic and Respiratory Society, Tokyo, 35-40, 2002 (in Japanese). clinical correlates. Ann Intern Med 119: 1198-1208, 1993. 12. Saito K, Nakayamada S, Nakano K, et al. Detection of Pneumo- 3. Hashimoto H, Kanai M. Side effects of glucocorticosteroids. Jpn J cystis carinii by DNA amplification in patients with connective Clin Exp Med 81: 798-802, 2004. tissue diseases: re-evaluation of clinical features of P. ca rin ii pneu- 4. Stuck AE, Minder CE, Frey FJ. Risk of infectious complications monia in rheumatic diseases. Rheumatology 43: 479-485, 2004. in patients taking glucocorticosteroids. Rev Infect Diss 11: 954- 13. Hughes WT, Kuhn S, Chaudhary S, et al. Successful chemopro- 963, 1989. phylaxis for Pneumocystis carinii pneumonitis. N Engl J Med 5. Wakefield AE, Pixley FJ, Banerji S, et al. Detection of Pneumo- 297: 1419-1426, 1977. cystis carinii with DNA amplification. Lancet 336: 451-453, 1990. 14. Sepkowitz KA, Brown AE, Armstrong D. Pneumocystis carinii 6. Takabayashi K, Kurasawa K, Yamazaki S, et al. Pulmonary oppor- pneumonia without acquired immunodeficiency syndrome: more tunistic infections in patients with autoimmune disease. J Jpn Soc patients, same risk. Arch Intern Med 155: 1125-1128, 1995. Intern Med 78: 1293-1298, 1989 (in Japanese). 15. Sepkowitz KA. Pneumocystis carinii pneumonia without acquired 7. Stringer JR, Beard CB, Miller RF, et al. A new name (Pneumocys- immunodeficiency syndrome: who should receive prophylaxis? tis jiroveci) for pneumocystis from humans. Emerg Infect Dis 8: Mayo Clin Proc 71: 102-103, 1996. 891-896, 2002. 16. Yasuoka A, Tachikawa N, Shimada K, Kimura S, Oka S. (1→3) 8. Beck JM, Warnock ML, Curtis JL, et al. Inflammatory responses beta-D-Glucan as a quantitative serological marker for Pneumo- to pneumocystis carinii in mice selectively depleted of helper T cystis carinii pneumonia. Clin Diagn Lab Immunol 3: 197-199, lymphocyte. Am J Respir Cell Mol Biol 5: 186-197, 1991. 1996. 9. Masur H, Ognibene FP, Yarchoan R, et al. CD4 counts as predic- 17. Okamoto K, Yamamoto T, Nonaka D, et al. Plasma (1→3)-beta-D- tors of opportunistic pneumonias in human immunodeficiency vi- glucan measurement and polymerase chain reaction on sputum as rus (HIV) infection. Ann Intern Med 111: 223-231, 1989. practical parameters in Pneumocystis carinii pneumonia. Intern 10. Gluck T, Geerdes-Fenge HF, Straub RH, et al. Pneumocystis Med 37: 618-621, 1998. carinii pneumonia as a complication of immunosuppressive ther-

Ⓒ 2008 The Japanese Society of Internal Medicine http://www.naika.or.jp/imindex.html

20